Intel Working to Boost Thunderbolt Speeds with Move to PCI-Express 3.0

[Stella]

You know I've been getting by "just fine" on FW800 for quite some time. The jump from FW800 to USB3 will be greatly significant. I am basing all of this on the fact that Apple will implement in on the new iMacs.

With that said, of course I will eventually migrate over to TB, but when an equivalent TB external cost 5-6 times more than its USB counterpart, something tells me to wait a little longer, which is exactly what I'll do.

But cool news either way!

USB3 support could be provided via an adapter. I'm surprised someone hasn't done this already. Not surprised Apple haven't ( they want to encourage people to buy TB instead - with only 1 or 2 expensive devices ).

Remembering Apple said "One cable to rule them all".

 

[MagnusVonMagnum]

So what will this mean for the first generation of Ivy Bridge Macs coming down the line? Are they going to be slower than ones later in the year or early next year in terms of Thunderbolt and/or the bus speed?

 

[Eidorian]

So what will this mean for the first generation of Ivy Bridge Macs coming down the line? Are they going to be slower than ones later in the year or early next year in terms of Thunderbolt and/or the bus speed?

I would not expect any updates to Thunderbolt speeds until 2014.

 

[WolfJLupus]

What he/she said...

...until you get more devices that use it then who really cares?

It sure sounds like they're having issues with the current technology to me since it's taking so long. I've been waiting for a decent enclosure and/or switch at a reasonable price for nearly a year now?

It all sounds great, external video cards, fast speeds, etc... Where's the hardware?! At least Apple could make something by now since third parties are way behind...

 

[wikus]

It all sounds great, external video cards, fast speeds, etc... Where's the hardware?! At least Apple could make something by now since third parties are way behind...

If Apple were to release any kind of device, you can be damn sure it would cost at least 50% more than anyone elses for NO LEGITIMATE REASON.

 

[MacOG728893]

If Apple were to release any kind of device, you can be damn sure it would cost at least 50% more than anyone elses for NO LEGITIMATE REASON.

That's sadly true and with the prices third parties are trying to sell at now, I couldn't imagine what Apple would try to charge.

 

[wikus]

That's sadly true and with the prices third parties are trying to sell at now, I couldn't imagine what Apple would try to charge.

The way things are going for Thunderbolt now and for over the last year, its headed in the same direction as Firewire.

Unfortunately for Apple theyre too damn psychotic with their pride. Sony had the right idea by making one of their Vaio laptops with Light Peak as a USB port, thereby allowing users to select WHICHEVER device suited them.

I've said it many times before and I'll say it again; Apple DOES NOT CARE about giving consumers *real* choice. Theyre only interested in giving you as much choice that will benefit THEM, not the user.

 

[MacOG728893]

The way things are going for Thunderbolt now and for over the last year, its headed in the same direction as Firewire.

Unfortunately for Apple theyre too damn psychotic with their pride. Sony had the right idea by making one of their Vaio laptops with Light Peak as a USB port, thereby allowing users to select WHICHEVER device suited them.

I've said it many times before and I'll say it again; Apple DOES NOT CARE about giving consumers *real* choice. Theyre only interested in giving you as much choice that will benefit THEM, not the user.

While I know this is true, I still obviously prefer Apple over most companies for computers and things related with tech. Some dude in another thread just keeps saying that it's pointless to buy USB3 when TB is so much faster.

His argument was that they are priced similarly for RAID arrays, but that is not what I am in the marker for.

I picked up 1TB CalDigit AV drive with FW800/USB3 for 179.99 and am loving it. Obviously I cannot use USB3 yet, but hopefully all that will change in the coming months. That's a lot cheaper than Lacie's 1TB little big disk TB external.

I think I'll be okay for now!

 

[backinblack875]

Technically you wouldn't want one, you would need one, because the device (formerly PC) you purchased was designed by the accounting department with the most crippled graphics on the planet.

Rocketman

lol thats what i figured. Would thunderbolt be powerful enough for that? that would be cool, if you need it.

also somewhat related question: would a thunderbolt external HD be fast to run a bootcamp windows partition on it?? (if possible at all anyways)

 

[linuxcooldude]

If Apple were to release any kind of device, you can be damn sure it would cost at least 50% more than anyone elses for NO LEGITIMATE REASON.

A lot of what Apple sells usually stays at the same price point since the last upgrade. So in some cases your getting more for the same price. With their software, prices have been going down drastically.

 

[MatthewAMEL]

going from 5 gigatransfers per second (GT/s) to 8 GT/s.

What did I miss? WTH is a 'gigatransfer'? How much data is in a 'transfer'?

 

[Umbongo]

going from 5 gigatransfers per second (GT/s) to 8 GT/s.

What did I miss? WTH is a 'gigatransfer'? How much data is in a 'transfer'?

5GT/s is 500MB/s and 8GT/s is 1GB/s.

http://en.wikipedia.org/wiki/PCI_Express
http://en.wikipedia.org/wiki/Gigatransfer

 

[deconstruct60]

Unfortunately for Apple theyre too damn psychotic with their pride. Sony had the right idea by making one of their Vaio laptops with Light Peak as a USB port, thereby allowing users to select WHICHEVER device suited them.

The USB standards committee nuked USB + TB combo solution. They disapproved it. That is a bit understandable since TB was in part an attempt to end-run them with a defacto USB 4.0. At this point USB still can roll out a fiber based upgrade in 4-6 years if they want and the cost structure for vendors makes sense.

Second, I don't think that was Sony's plan. I think Intel had prototypes of fiber USB 3.0 that they tweaked for TB. Sony's mistake was probably committing to the project before the stepping on the USB folks blew up and the miniDisplay port became the better option.

It actually would be cheaper for Sony to flip their proprietary port into a USB 3.0 one once the new support chipsets finish rolling out. They then can adopt the mini-DP like everyone else is doing to make their docking station work. As a laptop docking station as being a primary target the mini-DP port makes sense. Most laptops could use an external, larger screen about at least as much as an additional stationary GPU card. With a mini-DP port variant it does both.

----------

I would not expect any updates to Thunderbolt speeds until 2014.

Not just him.... Intel basically said as much
" ... As a result, Intel told me that we won't see any increase in Thunderbolt speeds for the next two years. ... "
http://www.anandtech.com/show/5405/the-first-thunderbolt-speed-bump-likely-in-2014

Too many changes too fast will impede the growth of TB. If folks want the costs to come down the volume has to go substantially up. Vendors aren't going to adopt it if it is a constantly moving target that requires design changes every 12-18 months. It just doesn't have that kind of inertia to justify that.

That said I wouldn't look for this PCI-e v3.0 bump at that two year mark. I'd bet that is 4-6 years out. The major problem is that they are going to come with something that is much faster than PCI-e v3.0 to transport it without inserting additional latencies into the datastream. That means TB will have to go faster than 3.0. Even the PCI-e folks aren't planning to go faster than v3.0 in the mainstream. v4.0 is going to be a special case short distance variant. It isn't going to be 'wake up tomorrow and flip the switch" issue for TB to jump past like a 4x v3.0 worth of bandwidth.

 

[Augure]

1. We basically don't care about TB upgrade for now, NONE OF US has any compatible devices because there aren't any available for consumer price and that's the problem

2. Which is a shame because, this could have been the future of cables ahead of USB3: it's thin, ultra fast in both direction, energy efficient and you can mount it in serie. I'm not only thinking about external portable hard drive but also about extern graphic GPUs that are going to be a big deal in the future, when ultra thin laptop are going to be nowhere to match the graphic processing power of the upcoming cards.

3. The problem, and we have to stop lying to ourselves (fortunately enough people on this topic seem smarter than the one on the Ipad topics), is once again the greed and pride of Apple: they have done everything to get the exclusivity to tackle competitors, they sell super expensive cables (50$ for a TB cable!!!), and they are working with no one to make it widespread.

Why would any manufacturer, whether PC or external devices, use it?

 

[KPOM]

Unfortunately for Apple theyre too damn psychotic with their pride. Sony had the right idea by making one of their Vaio laptops with Light Peak as a USB port, thereby allowing users to select WHICHEVER device suited them.

I've said it many times before and I'll say it again; Apple DOES NOT CARE about giving consumers *real* choice. Theyre only interested in giving you as much choice that will benefit THEM, not the user.

As others have pointed out, the USB Foundation nixed the use of the USB port. Apple had already open-sourced the mDP port, and so they went with that one.

Anyway, Thunderbolt has a good chance of becoming a standard port for things like docking stations. The Apple Thunderbolt Display is a good example of what is possible. Once Ivy Bridge comes out, Apple will support USB 3.0. The two standards have different uses. TB is overkill for flash drives. USB 3.0 is fine for that, but doesn't have the video capabilities of TB.

Also, Thunderbolt is primarily an Intel technology. Apple assisted in the design, but Intel is the main driving force. It will eventually be a part of the Ultrabook specification. So as much as we may mock MacBook Air clones, we should support their adoption, since it will help make Thunderbolt more useful.

 

[K2Climber]

Apple and Intel just announced a major update in thunderbolt technology.

This new update permits particulate matter to be transferred through the thunderbolt port at a speed so fast it can be sent into the future. In fact, one of the technicians working on the project was turned into particulate matter and sent into the future with a mission to bring back 3rd party devices we can use with the current thunderbolt port.

Unfortunately after sending him into the future, they realized they did not make the the 3rd party devices they need to bring him back from the future.

 

[macintoshi]

hehe

Yes, but not because of this. There's gonna be a major redesign this spring tough. The hard drive & DVD drive are gonna go, and it's gonna get alot thinner.

dream on

----------

I would not expect any updates to Thunderbolt speeds until 2014.

i don't believe apple will wait till 14 because its only a pci upgrade that shouldnt take that long seeing ivy bridge is still not out

 

[cube]

If the new Thunderbolt has DisplayPort 1.2 maybe it will be acceptable now.

Unless there is a faster DisplayPort coming out.

 

[bigwig]

USB 3.0 comes integrated into the Ivy Bridge chipset. Apple will be USB 3.0 soon with the next refresh of Macbook Air and Pro.

Which makes TB a dead letter as soon as that happens. Mac users will spend their dollars on USB3 peripherals they can buy rather than wait and hope somebody will produce affordable TB peripherals. Non-Mac users can't use TB peripherals anyway since their machines won't support it unless Intel requires it, which they (Intel) have already said they won't on mass-market hardware (which is where the volume is).

Critical mass matters. There isn't (and in my opinion never will be) a critical mass of available TB peripherals. As with Firewire 800, few non-Mac machines have it, so few manufacturers will produce for it. It's the polar opposite of a virtuous cycle.

 

[thekev]

Only you can decide that. Is 5 gigatransfers per second too slow for you?

Testing from anandtech in the past hasn't shown it to be a fully consistent number. They tested with displays as I recall, but some people will use displays on these machines that still have a single TB port.

I assure you it's more useful that the post you just laid down. Where else are you going to affordably get bi-directional 10Gbps throughput?

They have an interesting way of stating those numbers. Anyway it's not actually affordable on the device end, and don't blame the device manufacturers. It is currently a very tiny market with expensive hardware controllers, and Intel didn't really have a publicly available SDK for most of last year. You have to consider this. Any product release had to specifically target 2011 macbook pro, air, and imac users that would pay for such a device in its first generation. It's a niche market for now due to the inherent cost of entry. If TB sees reasonable adoption rates under Windows, you may start to see things improve.

I do not believe the concern is in the controllers on the system logicboard any longer. It is now on the added expense to the peripherals.

It'll still cost something to implement the thunderbolt chip which isn't inherently integrated, and passing it on to devices will not exactly bring prices down in the current year.

USB3 support could be provided via an adapter. I'm surprised someone hasn't done this already. Not surprised Apple haven't ( they want to encourage people to buy TB instead - with only 1 or 2 expensive devices ).

Remembering Apple said "One cable to rule them all".

I'm wondering if they still lack a stable driver.

I would not expect any updates to Thunderbolt speeds until 2014.

That's actually in line with Intel's initial estimate. They cited the 3 and 5 year marks, but obviously that could change.

If the new Thunderbolt has DisplayPort 1.2 maybe it will be acceptable now.

Unless there is a faster DisplayPort coming out.

I am not sure the mini displayport standard fully supports it. Apple says it's supported currently, but some features have never been seen under mini displayport. I'm not sure if it's a driver issue or hardware.

i don't believe apple will wait till 14 because its only a pci upgrade that shouldnt take that long seeing ivy bridge is still not out

Apple isn't the one who develops the chips. They gave it mini displayport which Apple licenses out with no fee, most likely to promote this connection on displays and third part accessories. Displays have way too much fragmentation at the moment. I wouldn't really mind it much if they actually released 10 bit drivers.

 

[macintoshi]

Apple isn't the one who develops the chips. They gave it mini displayport which Apple licenses out with no fee, most likely to promote this connection on displays and third part accessories. Displays have way too much fragmentation at the moment. I wouldn't really mind it much if they actually released 10 bit drivers.

Yes i know, thats why i said, ivy bridge is also still not out, so its quite possible for intel to push both: ivy bridge and the pci 3.0 to the new macbook pros

ps: my next macbook pro: ivy bridge i7 2.4ghz quad core, 2gb nvidia grafix, 512 or 1024gb ssd, 16gb 1333ghz ram, 16Gbps Lightpeak, this would make it for me

 

[Misaki]

lol thats what i figured. Would thunderbolt be powerful enough for that? that would be cool, if you need it.

also somewhat related question: would a thunderbolt external HD be fast to run a bootcamp windows partition on it?? (if possible at all anyways)

The problem with the TB is that a device hooked up to it is absolutely not the same as a PCIe 3.0 16 lane configuration, it's not even a PCIe2.0 16 lane configuration, it's a lot closer to 4 lanes at most... assuming there is no video going over it.

Storage is another story, it's traditionally been slow, and the fastest thing you can hook up to a SATA port is a ANS-9010 DDR2 RAM-Drive which is bottlenecked by the SATA interface itself at 500MB/sec. Any NAND or Spinning drive is going to be far slower. So 4 PCIe 2.0 lanes gives you 2000MB/sec.

A Video card on the other hand(AMD HD7970) can saturate the full 16GB/sec PCIe 3.0 16 lane configuration. You'd have to go back about 10 generations of video cards (back to AGP 8X) to find one that can't saturate 2GB/sec. Video cards have always been bottlenecked by their interface. So putting a (Primary) video card on the Thunderbolt port for video is just lunacy. When you see "USB Video card" what they mean is "additional non-accelerated screen", look up existing reviews for them, they largely complain about mouse latency and everything generally being worse then the onboard video above 1024x768.

At any rate, until Intel can figure out how to put the full 16 lanes of PCI 3.0 opticially to a separate device, there will be no "good" external video cards. Thunderbolt largely fills the gap needed for HD Video capture or playback (256MB/sec at 30fps 1080, 1GB/sec for 2K@60fps) to or from a display or a storage system.

 

[wizard]

Read the article, it won't be here anytime soon.

Two thousand pounds would be a lot to spend on a machine that could be out of date next month. :/

In any event the reason to wait for Ivy Bridge has little to do with PCI Express 3.0. It has more to do with the vastly improved integrated graphics, which is important for any machine where that is the only GPU in use.

In any event you are misguided here, every computer you will ever buy will be outdated upon leaving the store. It is the way it is. Some things are worth waiting for, like IB graphics, but that is only if you have a clear idea of when it is coming.

In this case the improved TB could come with IB (not likely), or it could come a year from now or even never. It is after all R&D. Just because a company works on something it doesn't mean it will ever be realized in a product.

 

[SactoGuy18]

I still think that one place where Thunderbolt ports will become important is theater-quality digital projectors.

Here's the reason why: it will allow a 4,000-line resolution digital movie to be stored on a single super-fast hard drive like a modified Western Digital Raptor 10,000 RPM drive in a single enclosure, with the movie data transferred from the drive to the projector using the Thunderbolt connection. So instead of needing a bank of drives for a single movie, the movie can be stored on a single drive, substantially reducing projector costs and also reducing the cost of making thousands of copies for digital projectors around the world.

 

[wizard]

It is still very early in the technology cycle.

I voted you up because in a sense I agree but you also need to understand it is very early in the technology adoption cycle. There is also a perception in the marketplace that Thunderbolt is a replacement for USB, it isn't. It will be a long time before hardware costs get down to around USB ranges.

I think Thunderbolt is a great spec. But until there are more peripheral devices to use with it, it really doesn't matter how fast it is...

Yep, specs are worthless without hardware implementing those specs. Just realize that such hardware does not come overnight. People need to think back to the first days of USB, Apple lead that revolution but it took years for USB to fully displace the other hardware of the time.

The existing PCIe 2.0 spec is lightning fast, if only we could get it showing up on docking stations, external hard drives, external graphics/physics cards, etc.

A few comments:

I believe Apples primary motivation was to implement TB to support docking monitors for its laptops. As you know that is where they implemented some of their first hardware. So from Apples standpoint they have already implemented their part.

Due to the inherent costs of TB I don't see it being cost effective for anything less than a disk array. Right now the TB cable itself is half the cost of a drive and then you need to add a TB specific adapter for the drive. So for the next couple of years one would expect most TB based "drives" to be disk arrays. An interesting exception here will be SSD, where there are now PCI Express controllers available to implement the SSD with.

As to external graphics or physics cards, sorry but I don't see the wisdom in such devices. TB simply isn't fast enough to justify the costs especially if you loose much of the cards performance because of that. Cards for computation might be viable but even then I don't see the value vs simply buying a Mac with a descrete GPU. To be viable external boxes for computation would have to be pretty high performance relative to what could be accomplished with a built in GPU. This relates to another thread I've taken part in, about compute nodes, the problem is how does one sustain a business on such hardware. Especially when the HPC sector is entrenched in the 1U compute server.

The economics of external compute devices is very difficult to get a handle on. For one thing you would need a big box just to house the power supply and fans. We are talking 300 watt and larger power supplies plus the GPU and support electronics. While it is easy to build such a box, it isn't so easy to stay in business selling them. It will be very interesting to see what the market does here.